4672 apparent position of the 0-0 peak by 1-2 nm to the red, so the true 0-0 position for the trans isomer is somewhat uncertain, and perhaps somewhat shorter than 632.5 nm.I2 The structure of the second series between 592.5 and 560.0 nm is not a reasonable looking progression for a single mode: either the apparent 1200-cm-' mode seen in the photoelectron spectrum is in fact a superposition of nearly degenerate modes, or there is some other more subtle effect leading to the odd looking intensity pattern of these peaks. Based on the normal-mode analysis of the neutral molecule,' the observed ion vibrational progressions are reasonable: only A, vibrations can be excited in this transition, and since it is a JI-system excitation, only carbon-framework motions are expected to be significantly excited. The ion progression at 350 cm-I corresponds to one of the two framework bending modes ~ 1 (453 2 cm-' in the neutral) and (347 cm-I), while the progression a t -1200 cm-' is associated with the framework stretching modes U S - U I O distributed from 1 187 to 1623 cm-l in the neutral. A Huckel-MO analysis14 of the I1 states involved in this optical transition yields additional qualitative insight: The principal bond length changes on excitation should be a lengthening of C2-C3 (and c 4 - C ~ ) and a shortening of C3-C4. Two modesI3 closely correspond to this motion: U S (1623 cm-') and v g (1238 cm-I). It is thus reasonable to suspect, as above, the simultaneous excitation of two framework-stretching vibrations in the ion. No similar elementary considerations seem appropriate to the bending motion, and we are unable to choose between the two likely modes. The photodissociation rate is linear in light intensity; it appears that the heat of dissociation to C&7+ H is less than the 2.0 eV corresponding to the 0-0 excitation, and that there is no reason to suspect a two-photon process.ls
+
Acknowledgment is made to the donors of the Petroleum Research Fund, administered by the American Chemical Society, and to the National Science Foundation for partial support of this research. W e wish to express appreciation to Professor J. I. Brauman for valuable comments about the interpretation.
Reactions of 5-Norbornene-2-diazonium Ions and Nortricyclene-3-diazonium Ions Sir:
The nature of the carbocations intervening in the acetolysis of 5-norbornen-2-yl brosylates (1-OBs) has not been established satisfactorily.' Both unsymmetrical (2a) and symmetrical delocalization (2c) have been postulated, based on contradicting reports on the scrambling of labels over positions 2,3 and 1,7.2-6 Readdition of acetic acid to norbornadiene as a source of scrambling has been maintained by one group6 but rejected by other^.^
1
2b
2a
24
W e report here on the decomposition of 5-norbornene-2diazonium ions (4) and nortricyclene-3-diazonium ions (8) in aqueous sodium hydroxide. Reactants and reaction conditions are such that 'carbocations will be generated and trapped irreversibly. The nature of the intermediates was investigated with the aid of optically active precursors as well as by deuterium labels. Photolysis of tosylhydrazone sodium salts generates diazo compounds' which are protonated by ROH to give diazonium ions and products derived therefrom.8 5-Norbornen-2-one tosylhydrazone (3) and nortricyclanone tosylhydrazone (7) afforded product distributions depending but slightly on the concentration of base (Scheme I, Table I). Optically active 39 produced 5 with 40-50% retention of configuration (50-60% racemization) (Table 11). Photolysis of 3 in D20-DONa introduced deuterium at C-2 of 4. Mass spectrometric analysiss,6 of [2Hl]-5indicated 27-28% rearrangement of D (Le., 54-56% equilibration), in agreement with the polarimetric data. These observations require an asymmetric precursor from which 40-50% of 5 is derived. If the asymmetric precursor was the homoallylic ion 2a, it should produce 6 with predominant
References and Notes (1) R. C. Dunbar in "Ion-Molecule Interactions", P. Ausloos. Ed.. Plenum Press, New York, N.Y., 1975, reviews this area. (2) R. C. Dunbar, Chem. Phys. Lett., 32, 508 (1975). (3) E, Fu, P. P. Dymerski, and R. C. Dunbar, J. Am. Chem. Soc., 98, 337 (1976) (4) 8. S. Freiser and J. L. Beauchamp. J. Am. Chem. SOC..96, 6260 (1974). ( 5 j M. T. Riggin and R. C. Dunbar, Chem. Phys. Lett., 31, 539 (1975). (6) J. R. Eyler, "An lntracavity Dye-Laser Technique for the Study of LaserInduced Ionic Processes", Presented at the 23rd Annual Conference on Mass Spectrometry and Allied Topics, Houston, Texas, May 25-30, 1975. (7) R. C. Dunbar, J. Am. Chem. Soc., 95, 619 (1973). (8) J. C. H. Hwa, P. L. de Benneville, and H. J. Sims, J. Am. Chem. Soc., 82, 2537 (1960). (9) M. Beez, G. Bieri. H. Bock, and E. Heilbronner, Helv. Chim. Acta, 56, 1028 (1973). Professor Heiibronner. and Dr. Bieri were kind enough to supply an expanded scan of the relevant spectral region from which the curve in Figure 1 was drawn, along with accurate values for the positions of the vibronic peaks. (10) R. Zahradnik and P. Carsky, J. Phys. Chem., 74, 1240 (1970). (11) P. P. Dymerski, E. Fu, and R. C. Dunbar, J. Am. Chem. Soc., 96, 4109 (1974). (12) Dr. J. P. Maier has been kind enough to give us preliminary data on optical emission from electron-impact excited hexatriene radical cation, which shows the 0-0 transition at 630.4 nm. Because of the possible distortion of our peak by a cis shoulder, we would take Maier's value as being a more precise absolute figure at present. (13) E. M. Popov and G. A. Kogan, Opt. Spektrosk., 17, 670 (1964)(English translation p 362). (14) Heilbronner and Straub. "Huckel Molecular Orbitals", Springer Veriag, 1966. (15) B. S. Freiser and J. L. Beauchamp, Chem. Phys. Lett., 35, 35 (1975). (16) Alfred P. Sloan Fellow, 1973-1975.
Robert C. DunbarI6 Department of Chemistry Case Western Reserce Unicersity Cleceland, Ohio 441 06 Receiued January 13, 1976 Journal of the American Chemical Society
/
98:15
/
Scheme I
I
H
'.NNHTS
7 Table I.
8
Product Distributions 5/6 Obtained from 4 and 8
Precursor
3
7
Scheme 111
0.1 h NaOH 0.2 U N a O H 0.5 'u h a O H 1.0 N U a O H
9.9190.1 10.3/89.7 10.8/89.2 11.4/88.6
3.9196.1 3.9196.1 4.4195 6 4.5195.5
4.0196.0 3.6196.4
J u l y 21, 1976
Table 11. Stereochemistry of the ( 1 q - 3
-+ 5
4673 6 Transformation
~~
6
5
Solvent
% O.D.
% D at C-1"
Config
% O.D.
1 N NaOH HzO-DME(312) 0.4 N NaOH HlO-CH30H (312) 0.3 N NaOH
41.4
27.6
S
2.1
27.0
R
3.2
R
1.8
49.6
Table 111. Product Distributions (%) from the Photolysis of 9 and 13 in 0.2 N NaOH
46.2
a Photolysis of 3 in DzO-DONa; estimated by mass spectrometry ( m / e 66.67) of the fragment containing C-1,4,5,6,7.
9 13"
10
11
12
73.8
24.5
64.6
17.9
1.o
(77.4)
(21.4)
(1.2)
15
14
1.7
-
-
7.6
7.7
1.2% of an unidentified product. Values in parentheses refer to
10
+ 11 + 12 = 100, for comparison with 9.
Scheme IV
Scheme I1 L C&OH,H+ 2 H6CH#X,SH 3. Raney-Ni
H&
0 (3R)" [aIz5D 74"(MeOH)
+&+AO
-s (3s)12
+
10
9
cH30&
[aFD - 49O(EtrOH)
12
11
L NaOH
2. 3. CF,CO,H CH,Li 4. LiAIH,
(3s)
[aYD
- 4O0(CHC13)
3 " 14
13
Scheme I11
HZ
(3s)
E
(3s)
a"D
- 67" (neat, 1dm) 0
NO
II I
H,NC-N-H
"exo" attack, i.e., ( S ) - 6from (1S)-3. W e established the absolute configuration and maximum rotation of 6 as shown in Scheme 11. The optical purity of 6 obtained from (1S)-3 was ]owl3 (Table 11). Moreover, the configuration of 6 varied with the solvent; e.g., in H 2 0 - D M E or in H20-CH30H it was opposite to that observed in water.I5 Consequently, we feel that the formation of 6 is not controlled by stereoelectronic factors such as homoallylic delocalization. W e attribute the residual optical activity of 6 to asymmetric solvation, and we take the stereochemical results as evidence against 2a. The optically active 5 may originate directly from the diazonium ion, particularly from endo- 4. Support for this view comes from the reactions of 8 and 13 (see below). It is seen from Tables I and I1 that the amounts of racemic (equilibrated) 5 obtained from 3 (6-6.5%) and 7 (4-4.5%) differ slightly. In order to estimate the fraction of nortricyclene-3-diazonium ions (8) which avoid formation of a "free" nortricyclyl cation, we generated (3S)-8 from a nitrosoamide p r e c ~ r s o r l(Scheme ~ . ~ ~ 111). 6 was obtained with 24% inversion of configuration. We may safely assume that the fraction of 8 reacting with inversion does not produce 5 (the sequence 8 2a 5 + 6 would afford 6 with retention). A single intermediate, the 3-nortricyclyl cation, then suffices to account for
- -
++
the formation of 6 and of racemic (equilibrated) 5 from the diazonium ions 4 and 8. By N M R spectroscopy in superacids the 3-nortricyclyl cation has been characterized as an essentially classical carbenium ion with slight charge delocalization into the cyclopropane ring.I8 The ratio of nucleophilic attack a t C-3 and C-1 (95:5) conforms to this model. Charge delocalization in the nortricyclyl cation becomes asymmetric and more intense upon introduction of a methyl group a t C-1. The product distribution obtained from 1methyl-3-nortricyclanone tosylhydrazone (9)19 (Table 111) reveals increased nucleophilic attack a t the substituted site, giving 2-methyl-5-norbornen-2-exo-ol ( l o ) , and a high ratio of exo:endo attack a t C-3 (11:12 = 14.5). No l-methyl-5norbornen-2-exo-o1(14) was found among the products from 9 whereas 7.6% of 14 arose from l-methyl-5-norbornen-2-one tosylhydrazone ( 13).21In addition, 13 afforded 4-methylbicyclo[3.l.l]hept-3-en-2-ol(15) by a C(l-7) migration.23It is tempting to attribute the formation of 14 and 15 to synchronous substitution and rearrangement of the endo-diazonium ion. The remaining products, 10-12, were formed in essentially the same proportions from 9 and 13. In summary, a rather simple mechanistic picture emerges from our studies on 5-norbornene-2-diazonium ions. A small portion (ca. 5%) of the products comes from an asymmetric, localized precursor which may well be the endo-diazonium ion. The majority of the products arise from the symmetrical, slightly delocalized nortricyclyl cation, in agreement with the solvolytic results of Cristol et aL5 There is no need for invoking asymmetric, delocalized intermediates (2a, b) and their interconversion. The variations induced by a 1-CH3 group are in accord with expectation. Communications to the Editor
4614
References and Notes (1) P. R. Story and B. C. Clark, Jr., “Carbonium Ions”, Vol. 111, G. A. Olah and P. v. R. Schleyer, Ed., Interscience, New York, N.Y., 1972, p 1016. (2) J. D. Roberts, C. C. Lee, and W. H. Saunders, Jr., J. Am. Chem. SOC.,77, 3034 (1955). (3) C. C. Lee and B. S. Hahn, J. Am. Chem. SOC., 92, 2583 (1970). I P. Vogel, R. Delseth, and D. Quarroz, Helv. Chim. Acta, 58, 508 (1975). ’ S. J. Cristol, T. C. Morrill, and R. A. Sanchez, J. Am. Chem. SOC.,88,3087 (1966); S. J. Cristol and D.A. Beimborn, ibid,, 95, 3651 (1973). C. C. Lee and E. C. F. KO,J. Am. Chem. SOC.,96,8032 (1974). W. G. Dauben and F. G. Willey, J. Am. Chem. SOC.,84, 1497 (1962). W. Kirmse and H. A. Rinkler. Justus Liebias Ann. Chem.. 707. 57 11967): see also W. Kirmse and R. Siegfried, 2. Am. Chem.’Soc.; 90; 6564 (1968). Repeated kinetic resolution with (R,R)-diethyl tartrate (incomplete formation of the acetal)’O provided a convenient access to optically active 5-norbornen-2-one. J. Y. Conan. A. Natat. F. Guinot, and G. Lamaty, Bull. SOC.Chim. Fr., 1400, 1405 (1974). J. S. Bindra, A. Grodski, T. K. Schaaf, and E. J. Corey, J. Am. Chem. SOC., 95, 7522 (1973). (a) J. A. Berson and R. G. Bergman. J. Am. Chem. SOC.,89, 2569 (1967); (b) J. A. Berson, R . G. Bergman, J. H. Hammons, and A. W. McRowe, ibid., 89, 2581, (1967); (c) J. A. Berson, R. G. Bergman, G. M. Clarke, and D. Wege, bd. 91, 5601 (1969). The maximum rotation [ a I z 540.1’ ~ (EtOH), given in ref 12c is somewhat lower than ours, due to the uncertain optical purity of a reference compound. In addition to 5 and 8, the photolysis of 3 affords minor amounts of unsaturated isomers, e.g., bicycl0[3.l.l]hept-3-en-2-o1,’~ which are not readily separated from 8. Therefore, 6 was purified by ozonolysis, followed by GLC and sublimation, before rotations were taken at 365 nm; maximum [ a I z 5 3 6 5 116.6’ ( C 21.4, CHC13). W. Kirmse and R . Sieofried. Chem. Ber.. 105, 2754 (1972). (15) Photolysis of (1S t 3 inH20-CH30H (3:2) afforded 3-methoxynortricyclene (40%) and 2-methoxy-5-norbornene (7%) in addition to 5 (10%) and 8 (39%). The configuration of 3-methoxynortricyclene was (S). opposite to that of 6 obtained in the same run. (S)-3-Methoxynortricyclene. prepared by methylation of (S)-6, had [ c Y ] ~ 1.48’ ~D (neat, l = 1 dm). (16) Some leading references concerning the basic hydrolysis of nitrosoamides are given below: (a) R. A. Moss, J. Org. Chem., 31, 1082 (1966); (b) W. Kirmse and G. Wachtersbuser, Justus Liebigs Ann. Cbem., 707,44 (1967); (c) R. A. Moss, D. W. Reger, and E. M. Emery, J. Am. Clem. Soc.,92, 1366 (1970); (d) S. M. Hecht and J. W. Kozarich, J. Org. Chem., 38, 1821 (1973). (17) No incorporation of deuterium into 6 was observed when 8 was treated with 0.1 N NaOD in DzO. This excludes intervention of Bdiazonortricyclene as a possible source of racemization. (18) G. A. Olah and G. Liang, J. Am. Chem. SOC.,95,3792 (1973): 97, 1920 (1975). (19) Mp 134-135 OC, obtained from I-methyl-3-nortricyclanonezo and tosylhydrazine in methanol-HCI. (20) J. Paasivirta, Ann. Acad. Sci. Fenn., Ser. A2, 116, 8 (1962). (21) Mp 140-141 OC, obtained from l-methyl-5-norbornen-2-oneZz and tosylhydrazine in methanol-HCI. (22) H. L. Goering and C. S. Chang, J. Org. Chem., 40, 2565 (1975). (23) The C(1,7) shift sould lead to alcohols 15 and 15’ (R = H) by way of the allylic cation 16. 15’ (R = H) was not observed and may have rearranged during GLC. However, photolysis of 13 in CH30H-CH30Na afforded both 15 and 15’ (R = CH3) in 4-5% yield.
13-
S p e ~ t r a land ~ ? ~~ h e m i c a l ~evidence ,’~ are in good agreement with a “contact pair” (“contact species”) strukture” and rule out the presence of (solvent-separated) ion pairs or even free carbanions in substantial concentrations. 2-Isomers ((2)-1) may be converted into the corresponding E-isomers ((E)-1) and vice versa, as was previously established for GrignardI2 and lithiumI3 compounds. New and astonishing, however, is the o b s e r ~ a t i o n that ’ ~ ~allylpotassium ~~~~~ derivatives undergo such configurational changes only very slowly,14 half-lives of pure isomers falling in the range of hours a t normal temperatures in the case of simple hydrocarbon derivatives. H
H\
c= /
R
H I /C
/ C\
CH3
1
R
rapid
R
H
rapid
-
H H \ / /c=C \
CH3
= C\ /
1
H
slow
CH2K / C\
\
,c=
CHzK
R
H
Most surprising was the finding that butenyl-,’ h e ~ e n y l - , ~ 2-methyl-b~tenyl-,’~and other alkenylpotassium compounds1~5~16 drastically favor the Z-configuration. Although base-catalyzed alkene isomerizationsI7 have been known to be highly cis selective, this does not necessarily imply an exceptional thermodynamic stability of the intermediary endo(cis)-alkenyl carbanion, since these isomerizations occur under kinetic control and thus may.merely reflect differences in the hydrogen transfer mechanisms for cis- and trans-alkene formation.l8,I9Furthermore the Z preference of alkenylpotassium compounds may be, but do not have to be, related to the strong tendency of l-phenylbutenylpotassium20to accommodate the terminal methyl group in an endo(cis) position. In liquid ammonia,20 the phenylbutenylpotassium probably forms a solvent-separated ion pair and it benefits from extensive charge delocalization. Therefore, fundamental dissimilarities with alkenylmetal compounds should still be envisaged. For instance, the electron excess of the 1-phenylbutenyl anion could mainly reside in the aromatic ring and thus make the carbanion (limiting structure 2) resemble the Z-crotonitrile (3) which is known to be more stable than the corresponding E-isomer.21 In order to gain deeper insight into the stereochemical behavior of alkenylmetal compounds, we investigated the effects of the organic part, the metal, and the solvent on the equilibrium position.
d+R&+ /
/
16
OR 15
15‘
Wolfgang Kirmse,* Norbert Knopfel Abteilung fur Chemie der Ruhr-Universitat 4630 Bochum, Germany Received April 5, 1976
H\
/C
/H
= C\
H3C
2-Alkenyl Anions and Their Surprising Endo Preference. Facile and Extreme Stereocontrol over Carbon-Carbon Linking Reactions with Organometallics of the Allyl Type Sir: Allyl type organometallic compounds merit special attention on account of their potential in the field of organic synthesis.] Of particular interest is their stereoselective and regioselective behavior. Organometallic compounds of the allyl type, previously hardly known,2 became easily available through alkene metalation with butyllithium in the presence of potassium tert-butoxide3-‘ or with trimethylsilylmethylpotassium.6~7 Journal of the American Chemical Society
98:15
CH
H\
IH
/c= H3C
C\ C \\I
\\
N. 3 2
In general, the Z/E equilibrium for each system was established by the metalation of each of the pure olefin isomers. The resulting reaction mixture was stirred until the equilibrium point was reached (at least 24 h in tetrahydrofuran”), oxirane was then added and the derived isomeric alcohols were analyzed by gas phase ~ h r o m a t o g r a p h y The . ~ ~ Z/E ratios of the product mixtures should be equal to the Z / E ratios of the organometallic precursors provided that product yields are high and a/y-ratios (oxirane attack a t the vinylogous vs. the terminal position of the allylic system) observed for both confi-
/ July 21. I976
